SASVAAI/qwen38-27b-terraform

VISIONPricing:Input $1.6 / Cached $0.15 / Output $12Concurrent Unit Cost:2Model Size:27BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 23, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

SASVAAI/qwen38-27b-terraform is a 27.78 billion parameter merged fine-tune of Qwen/Qwen3.8-27B, developed by SASVA AI Model Cognition Labs (MCL) Team. This model is specifically optimized for generating Terraform (HCL) configurations from plain-language descriptions, supporting AWS, GCP, Azure, Kubernetes, and eleven other provider families. It was trained with an 8,192 token context length and excels at infrastructure-as-code generation.

Loading preview...

Overview

SASVAAI/qwen38-27b-terraform is a specialized 27.78 billion parameter language model developed by SASVA AI Model Cognition Labs (MCL) Team. It is a merged fine-tune of the Qwen/Qwen3.8-27B base model, meaning its LoRA adapter has been fused into the base weights, allowing it to load as a standard transformers checkpoint without PEFT dependencies. The model's primary function is to generate Terraform (HCL) code from natural language descriptions, supporting a wide range of cloud providers and infrastructure types.

Key Capabilities

  • Terraform HCL Generation: Translates plain-language infrastructure requirements into complete and valid Terraform HCL configurations, including resource, data, variable, output, provider, and terraform blocks.
  • Multi-Provider Support: Fine-tuned on data covering AWS, Google Cloud, Azure, Kubernetes, OCI, Alibaba Cloud, GitHub, DigitalOcean, IBM Cloud, Yandex, Vault, Docker, OpenStack, vSphere, and Cloudflare.
  • Optimized for Code Generation: Trained with a specific prompt format that includes a system prompt defining it as a "Terraform expert" and a user prompt with a fenced code block for the description.
  • Efficient Deployment: Available as unquantized bfloat16 safetensors for transformers and as a Q4_K_M GGUF for local inference with Ollama, offering flexibility in deployment.

Good For

  • Infrastructure-as-Code (IaC) Automation: Developers and DevOps engineers looking to rapidly draft Terraform configurations based on high-level descriptions.
  • Reducing Manual HCL Writing: Automating the creation of .tf files for various cloud and infrastructure providers.
  • Prototyping and Learning: Quickly generating examples of Terraform code for new projects or educational purposes.

It's important to note that the model was trained with a specific ChatML template and system prompt; deviating from this format may lead to suboptimal performance. The model's context length during training was 8,192 tokens, though the base model has a capability of 262,144 tokens.

Finetunes
1 models